[ridaj]

I want to believe that the future is decentralized but I'm convinced of the opposite when it comes to online media, primarily because spam, abuse and content moderation are expensive functions to perform in which there are strong scale effects, and they are critical not just to ad-based monetization but perhaps also to broad societal acceptance of technology.

[JohnBooty]

Yeah. Unfortunately... yeah.

The dream of decentralized everything sounds wonderful, but there's an implicit assumption that most folks involved are good actors.

It quickly becomes untenable when corporate-backed or state-backed bad actors are introduced.

Here's a very simple thought exercise for anybody who disagrees. Imagine a modest social media team of perhaps 10 paid employees. That's 24,000 people-hours' worth of content generation per year... and it is perhaps multiplied by a factor of ten if they're sophisticated enough to put some work into tooling to automate their work. Russia alone reportedly had hundreds of people doing this sort of work, and surely they're not alone.

Now imagine 10,000 of those teams around the world. That's 240,000,000 people-hours of span and/or bad-faith social media posting. And that's probably an extremely conservative estimate.

How would a decentralized social network possibly combat this?

It's a life-or-death struggle even for a company with deep pockets (and huge cubical farms full of human content moderators) like Facebook to combat this sort of thing.

If one of the decentralized solutions ever reaches any sort of critical mass, it will have to confront this in a hurry, and it will not be able to.

[feanaro]

The same analysis makes it untenable for centralized platforms, as you noticed yourself.

It's arguably harder to do it on a decentralized platforms since they are... Well, decentralized. Those 240,000,000 people hours are then spread over all of the decentralized platforms instead of being focused on one or a few. Also, the cost of moderation is spread over everyone instead of a single entity having to pay for all of it.

[JohnBooty]

    The same analysis makes it untenable for centralized platforms, as you noticed yourself.

Strong disagree, though I certainly hope that I am wrong and that decentralized platforms are more viable than I suspect.

It is not easy, but it is most definitely tenable for centralized platforms with sufficient funding and motivation i.e. Facebook, Google, etc.

They have:

1. Many millions (billions, in Facebook's case) of users, some percentage of which are willing to click those "Report..." buttons

2. Their own farms of in-house workers

3. Their own bespoke tools and heuristics to identify trends and more easily identify malicious actors/content, for review by the human moderators. this would presumably incorporate reports from users.

    It's arguably harder to do it on a decentralized platforms

Again, I really can't agree but hope to be wrong.

How does a malicious social media team "attack" Facebook? For the most part, they produce content that winds up being spread virally. They don't spead the information around Facebook; the users do. This would function largely the same way in a decentralized network.

Keep in mind that today's decentralized networks are a bit like the Usenet was before "Eternal September."

https://en.wikipedia.org/wiki/Eternal_September

In other words, these decentralized networks are currently populated primarily with relatively savvy users. Not perfect by any means, but on average, more savvy than your average FB or Twitter user who will eagerly share articles and links without even a cursory vetting. That will change in a hurry if any of these decentralized networks ever gains critical mainstream mass.

[feanaro]

To me it seems that ad-based monetization is untenable. The more obnoxious it becomes, the more people are enticed to learn about and set up adblockers.

[sverhagen]

Maybe they mean, "a better future _would_ be distributed"...